Project Summary Millions of Americans suffer from migraine headache attacks, but the causes of these individual attacks are not known. While many factors, or ?triggers?, have been thought to cause migraines (e.g., certain foods, weather, stress), the influence of these diverse variables is difficult to measure or quantify. Currently, health care providers, individual migraine sufferers, and researchers are unsure which factors actually are associated with migraine and what can be done to reduce the risk of migraine for any individual. Focusing on the variability in these triggers and quantifying this variability using information theory allows the potential influence of all triggers to be measured on the same scale. This project will create a trigger surprise model that allows all types of migraine triggers to be accurately measured for use in predicting risk of migraine attacks. In Aim 1, a longitudinal cohort study design will be used to observe N = 200 individuals with migraine to examine the natural variability in migraine activity and many common migraine triggers. In Aim 2, a measurement system will be created that utilizes the within-person probability distribution for each trigger to estimate a daily risk score for a migraine attack. This measurement system will be immediately valuable in reducing the current confusion surrounding migraine triggers, could be used in forecasting future attacks, and would be crucial to the further study of the biological mechanisms underlying the initiation of a migraine attack.